This is a good example of a problem that's easy to talk about but difficult to fix:
"tell CPU to sleep!" versus this whole lengthy article.
Wondering now if the issues with Linux laptop sleep are related or have similarity.
> The furi_hal_os_sleep() function performs tick suppression inside the OS domain.
Reminds me that many eons ago I witnessed the introduction of CONFIG_NO_HZ in Linux, is it the default now? Combined with timer coalescing this makes for big energy savings.
Can you provide more detail?
I'd be happy to return the favor with my use of NO_HZ and NO_HZ_FULL!
https://arstechnica.com/gadgets/2013/06/how-os-x-mavericks-w...
Generally the hardest part is simply making sure that the device can wake up correctly and quickly enough under all situations. Also, tracking down unexpected power draws gets into a nice crossover between electrical and firmware design teams which can either be great or extremely frustrating for either side of that interaction.
I know it's been a problem on MacOS forever, like back in the 68k days. Some peripherals never come back from sleep and you have to power cycle them to get them to "wake up." And sometimes MacOS doesn't come back either.